Antiscatter Grids - An Essential Component of Image Quality - Part 2
by Dr. Frank Ranallo, PhD
n Part I of this series of articles on grids I described five
very important characteristics of grids. As a review they are: 1) the
orientation or direction of the grid lines, 2) the grid focal distance or focal
range expressed in inches or cm, 3) the grid ratio (which may range from 1:4 to
1:17), 4) the grid line frequency expressed in lines per inch or lines per cm,
and 5) the thickness of the lead septa – which is usually not indicated on the
grid. It was also noted that many grids for upright use are labeled as 40 to 72
inch grids or are used at both 40 and 72 inches, which will result in inferior
imaging at either 40 or 72 inches or both. However there is one important twist
to this problem not mentioned in Part 1 that now will be discussed.
We have always been taught that the grid ratio is the prime
specification that controls both scatter cleanup and the carefulness with which
we need to align the grid. However this is not necessarily true for present day
grids. It is not the grid ratio alone, but the combination of the grid ratio
and the lead septa thickness that determines (1) the degree of scatter rejection
and contrast improvement, (2) the acceptable amount of lateral decentering of
the grid, and (3) the allowable range of usable focal distances. This fact has
been demonstrated by the availability of some high quality grids that have grid
ratios as high as 17:1, in applications where common wisdom would say that the
grid ratio should not exceed 10:1. Yet these 17:1 grids can perform very well,
better than a more conventional grid, as we will see below.
What is the affect of varying the lead septa
thickness in a grid? The diagram below shows some of the effects and
compromises.
FIGURE 1
The two top grids have the same grid ratio.
However the top grid has septa which are twice as thick as the middle grid. If
these were 10:1 grids and the septa thickness of the top grid was 40 microns
(0.040 mm), then the one scattered x-ray shown must actually pass through about
0.40 mm of lead -- the lead septa thickness times the grid ratio of 10. This
amount of lead provides substantial attenuation to the scattered radiation,
letting only about 2 to 5% of the radiation through. The lead septa of the
middle grid have a thickness of 20 microns. Scattered x-ray like those on the
left that pass through only one septa must pass through about 0.20 mm of lead.
This allows about 10 to 25% of the radiation through, and is much less effective
at attenuating the scattered radiation. Scattered radiation passing through the
grid at a larger angle, like the one shown on the right, will pass through 2
septa and will be attenuated as much as the x-ray shown in the top grid. Even
though the two top grids have the same grid ratio, the middle grid will have
poorer scatter cleanup than the top grid.
The scatter cleanup of the middle grid will
be only a bit better than that of the bottom grid (with a grid ratio of 5:1):
The scattered radiation shown on the left will be unattenuated but the scattered
radiation on the right will be attenuated to the same amount as in the middle
grid. By reducing the septa thickness of the middle grid, it "behaves" more
like a 5:1 grid than like a 10:1 grid.
The advantage of thinner septa (as in the
middle grid) is better transmission of the primary beam, since less of it is
blocked by the thinner septa. This is illustrated below.
FIGURE 2
These two grids are similar to the two top
grids in the previous figure. The only difference is an increase in the grid
line frequency in both grids, perhaps from 102 lines/inch to 173 lines/inch.
Here you can see that the thicker septa in the top grid
will cause substantial attenuation of the primary beam, more so than the second
grid. Thus as you go to very fine line grids you are usually forced to reduce
the septa thickness and will need a higher ratio grid to have the same amount of
scatter cleanup.
Let's now look again at the grid that I
criticized in Part 1 of this series: a 12:1 or 10:1 grid that had a labeled
focal range of 40 to 72 inches. I had said that this large a focal range was
not possible for a 10:1 or 12:1 grid, and that is true if the normal assumption
is made that the grid septa are highly attenuating to anything other than
properly aligned primary radiation. However if the lead septa are thin, then
the intensity falloff that occurs is reduced when the focal spot is not near the
grid focus. More of the misaligned primary radiation will be able to penetrate
the grid, just like the scatter in Figure 1 that only passed through one thin
septa.
So you may think that you have a 12:1 grid
that appears to work at both 40 and 72 inches. And yes, it technically may be a
12:1 grid, and there may not be significant artifacts at either 40 or 72
inches. However it then must have rather thin lead septa which give it the
scatter cleanup that is more like a conventional 6:1 grid. This will degrade
the image contrast and image quality. You may think that if you have CR you
could compensate for the loss of image contrast, and you can, but the overall
image quality will still be degraded. This is due to a reduction in the ratio
of contrast to image noise, particularly in areas of high patient attenuation.
Some equipment manufacturers can use the
above ideas to their advantage. When needing a high frequency grid, they
intentionally reduce the septa thickness to allow more primary transmission and
then increase the grid ratio to provide proper scatter cleanup. The result is
the good 17:1 grids mentioned earlier.
Now that I have discussed grids and their
importance in image quality, in Part 3, the last part of this series about
grids, I will discuss some of the important tests that should be performed with
grids to verify their quality and proper alignment.
Dr Frank Ranallo,
PhD, DACR is Associate Professor of Medical Physics and Radiology at the
University of Wisconsin Medical School in Madison, Wisconsin.
Questions or
comments concerning this article will be forwarded to Dr. Ranallo if sent to the
following email address:
rls@gammex.com
| 
